Geoscience Reference
In-Depth Information
Combining Two Clustering Ideas
for Typification of Ditches
Jing Tian, Wen-Yu Yang and Li-Jun Chen
Abstract Ditches are important components of hydrological ecosystem and play a
critical role in capturing and removing the micro-organisms and pollutants from
water body. The generalization of ditches supports the multi-scale analysis and
modeling of hydrological and ecological environment that relate to ditches. This
paper proposes and implements a new method that combines two clustering ideas
for typification of ditches. This method groups the ditches based on the edge-
cutting of the minimal spanning tree of ditches, and represents the ditches in each
group based on K-means++ algorithm. This study also presents the validation of
the method with the experimental data of Guangzhou city.
Keywords Map generalization
Typification
Ditches
Clustering
1 Introduction
Ditches are important components of hydrological ecosystem and have a great
influence on capturing and removing the microorganisms and pollutants from
water body (Dunn and Mackay
1996
; Bennett and Moore
2005
; Smith
2009
). The
cartographic generalization and multi-scale representation of ditches support the
multi-scale analysis, as well as the modeling of hydrological and ecological
environment. A simple selection algorithm is not enough to maintain the distri-
bution pattern of ditches that are approximately parallel distributed (Sandro et al.
2011
). Therefore, typification is needed to address this issue.
Typification is a generally accepted operator in the community of map gener-
alization (Shea and McMaster
1989
;Li
2007
). It reduces the quantity of the objects
